|
Keywords
|
Aerosol pollution, Wetland degradation, Remote sensing, Climatic factors, Human activities,
Machine learning
|
|
Abstract
|
This paper investigates the spatiotemporal patterns of aerosol optical depth (AOD) over 25 wetlands
and lakes in Iran during the period 2000–2024. Monthly AOD values were integrated with climate and
terrestrial variables. Trend analysis was performed using the Hamed-Rao Mann-Kendall test and Sen’s
slope estimator. The results indicated that AOD increased by 8 and 12% over Iran’s water bodies in
winter and autumn, respectively, while no significant increasing trends were observed in spring and
summer. Annually, 8% of Iranian wetlands faced a significant increasing trend in AOD, which mostly
occurred over the Jazmourian and Sheedvar wetlands (Z > 2.5). Seasonally, a notable increasing trend
occurred over the Amirkelayeh, Parishan, and Mighan wetlands, as well as Lake Urmia (Z > 3). Using
Random Forest and XGBoost models combined with SHAP analysis, it was determined that AOD
changes were primarily associated with wind speed and vegetation cover in Jazmourian, and with
salinity and soil moisture in the Parishan wetland. Terrestrial factors (80.3%) had a greater contribution
to the degradation of air quality in both wetlands compared to climatic factors (19.7%). These findings
are useful for understanding the synergy between natural and anthropogenic factors governing air
quality in aquatic ecosystems in Iran
|